基于机器视觉PCB裸板缺陷自动检测研究

运用参考比较法对机器视觉PCB裸板缺陷检测进行了研究。在相机摄像头下同一位置采集多幅标准PCB图像累加求平均值得到标准电路板图像,运用Harris角点算法进行标准电路板图像和待测电路板图像的配准,分别对标准电路板图像和待测电路板图像进行灰度变换、中值滤波、二值化、异或等图像处理检测出缺陷区域,然后通过形态学消除伪缺陷,实验证明,该检测方法有较高的准确率。

Automatic Defect Inspection of PCB Bare Board Based on Machine Vision

Applying to the method of reference comparison to automatic defect inspection of PCB bare board based on machine vision.Camera captures multiple standard PCB image and calculate its average gray get standard circuit board image tn the same position,image registration completed by standard PCB image under test PCB image’s corner detection and corner registration,adopting to standard PCB image under test PCB image adopt gray-scale transformation,filtering,binarization,XOR and other image processing respectively to detect the position of the defect area.